Shoes++
Shoes++
Shoes++
“Shoes++: A Smart Detachable Sole for Social Foot-to-foot Interaction”. In Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT’22).
“Shoes++: A Smart Detachable Sole for Social Foot-to-foot Interaction”. In Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT’22).
“Shoes++: A Smart Detachable Sole for Social Foot-to-foot Interaction”. In Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies (IMWUT’22).
Introduction
Introduction
Introduction
Foot is an important body part that not only performs locomotion but also participates in intent and emotion expression. Thus, foot gestures are an intuitive and natural form of expression for interpersonal interaction. Recent studies have mostly introduced smart shoes as personal gadgets, while foot gestures used in multi-person foot interactions in social scenarios remain largely unexplored. We present Shoes++, which includes an inertial measurement unit (IMU)-mounted sole and an input vocabulary of social foot-to-foot gestures to support foot-based interaction.
The gesture vocabulary is derived and condensed by a set of gestures elicited from a participatory design session with 12 users. We implement a machine learning model in Shoes++ which can recognize two-person and three-person social foot-to-foot gestures with 94.3% and 96.6% accuracies (N=18). In addition, the sole is designed to easily attach to and detach from various daily shoes to support comfortable social foot interaction without taking off the shoes.
Based on users' qualitative feedback, we also found that Shoes++ can support team collaboration and enhance emotion expression, thus making social interactions or interpersonal dynamics more engaging in an expanded design space.
Foot is an important body part that not only performs locomotion but also participates in intent and emotion expression. Thus, foot gestures are an intuitive and natural form of expression for interpersonal interaction. Recent studies have mostly introduced smart shoes as personal gadgets, while foot gestures used in multi-person foot interactions in social scenarios remain largely unexplored. We present Shoes++, which includes an inertial measurement unit (IMU)-mounted sole and an input vocabulary of social foot-to-foot gestures to support foot-based interaction.
The gesture vocabulary is derived and condensed by a set of gestures elicited from a participatory design session with 12 users. We implement a machine learning model in Shoes++ which can recognize two-person and three-person social foot-to-foot gestures with 94.3% and 96.6% accuracies (N=18). In addition, the sole is designed to easily attach to and detach from various daily shoes to support comfortable social foot interaction without taking off the shoes.
Based on users' qualitative feedback, we also found that Shoes++ can support team collaboration and enhance emotion expression, thus making social interactions or interpersonal dynamics more engaging in an expanded design space.
Foot is an important body part that not only performs locomotion but also participates in intent and emotion expression. Thus, foot gestures are an intuitive and natural form of expression for interpersonal interaction. Recent studies have mostly introduced smart shoes as personal gadgets, while foot gestures used in multi-person foot interactions in social scenarios remain largely unexplored. We present Shoes++, which includes an inertial measurement unit (IMU)-mounted sole and an input vocabulary of social foot-to-foot gestures to support foot-based interaction.
The gesture vocabulary is derived and condensed by a set of gestures elicited from a participatory design session with 12 users. We implement a machine learning model in Shoes++ which can recognize two-person and three-person social foot-to-foot gestures with 94.3% and 96.6% accuracies (N=18). In addition, the sole is designed to easily attach to and detach from various daily shoes to support comfortable social foot interaction without taking off the shoes.
Based on users' qualitative feedback, we also found that Shoes++ can support team collaboration and enhance emotion expression, thus making social interactions or interpersonal dynamics more engaging in an expanded design space.
My Contributions: led the participatory design workshop; Summarized input vocabulary; Designed and developed the sole.
My Contributions: led the participatory design workshop; Summarized input vocabulary; Designed and developed the sole.
My Contributions: led the participatory design workshop; Summarized input vocabulary; Designed and developed the sole.
Learning & reflection
Learning & reflection
Learning & reflection
Challenges and Opportunities for Practical Uses
Practical Design Consideration: For a sole to be practically usable, the current design needs to be improved by waterproofing the electronic components and reducing the overall volume of the system. Firstly, while the material for construction is waterproof, we did not take into account the structural design. In the future, we can further waterproof the system by completely insulating the PCB and all electrical connectors with silicone gel during fabrication. Secondly, the current design carries all IMUs, PCBs, and batteries separately on the elastic sole. Such design of the external soles might be problematic, because of their thickness in some applications. For example, in professional dance learning, dance shoes typically have very thin leather soles. Although the sole might be comfortable enough for normal dancing, it probably requires a modification for the professional dance movement. Future prototypes can miniaturize the whole design by integrating all components into one button-like gadget for more versatile and adaptive applications.
Power Management: Since the MEMS IMUs we used do not consume a lot of power, we can use energy harvesting methods such as near-field communication (NFC) coil to power the electronics [1]. NFC is a wireless technology capable of transmitting power between integrated components through inductive coupling between coils. The transmitter generates an oscillating electromagnetic field, which is then picked up by the receiver to obtain power. Some low-power NFC devices known as NFC tags can be battery-free and economical, yet only capable of working over short distances. Since NFC devices are also typically made of silicone [2], we can seamlessly integrate the NFC coil into the substrate itself as part of the system by using the same material, increasing the overall flexibility and robustness of the wireless system. When combined with an energy collector, the NFC power supply system can capture data when the label is not powered on. Furthermore, we can also implement other methods of battery-free communication and energy collection, such as the ambient backscatter technology [3], the piezoelectric method to collect energy from vibration [4], or the thermoelectric effect to collect energy from heat transfer [5].
Activation Trigger and Auto Segmentation: We envision social foot-to-foot input as a complementary and alternative input method in a variety of future collaborative computing device ecosystems. In real life, the system needs to distinguish between intentional interactions with the device and unintentional daily activities, including walking, running, and so on. To avoid accidental triggering, it can only respond after activation. Users can choose between an "activation pose" and a "termination pose" that contain a special command (or series of commands). In addition, sometimes participants had to enter multiple gestures in sequence in the same context. Currently, our gesture segmentation is based on the difference between resting and moving, but in practice, the user may perform gestures continuously. In this case, we plan on using signal processing or machine learning models for gesture detection and partition to separate users' gestures. Further processing and updates of these algorithms will enable the use of Shoes++ in a real-world setting.
Opportunities for Scalability
Create Gesture Vocabulary with More People: Currently, due to the limited number of devices and capacity of the experiments, we only designed social foot-to-foot gestures based on two and three participants as a trial. We believe the concept of interactive foot gesturing, with a goal to complete tasks collaboratively, is valid as we move into larger groups of participants. Each additional participant will bring an extra layer of complexity when analyzing gesture semantics, and we plan to include more multi-user foot gestures to complete the vocabulary. Moreover, the spatial variations and user experience in larger group-based interactions can also be further explored.
Front-end Software for Shoes++ Design: In our post-study interviews, many participants expressed an interest in the aesthetic customization of Shoes++. They mentioned that they would be willing to wear the device in their daily lives if they have the opportunity to choose from a variety of styles, patterns, materials, and colors. Currently, our Shoes++ is more focused on structural versatility, but in the future, we can improve the appearance design. For example, we can provide a front-end software and appearance rendering tool to help users design their own Shoes++.
Using Synthetic Data to Enlarge Training Set: In real-world applications, collecting training data from users, especially if it is collected from multiple users, is expensive and time-consuming. To solve this problem, we plan to use a synthetic non-word approach to extend the training set. This approach requires a 2D/3D human body gestures simulation system [6] that continuously generates human body models capable of making dynamic foot gestures. We can choose the location of the IMUs on the foot of the 3D human body and calculate the data of virtual IMUs from motion data. A series of works in computer vision has demonstrated the feasibility of controlling 2D/3D human bodies and using virtual IMUs. For example, Kwon et al. proposed IMUTube, which retrieves videos from public repositories and subsequently generates virtual IMU data from the videos [7]. Through these studies, we believe it may be possible to synthesize datasets of various foot movements in multiple people to facilitate future training.
Challenges and Opportunities for Practical Uses
Practical Design Consideration: For a sole to be practically usable, the current design needs to be improved by waterproofing the electronic components and reducing the overall volume of the system. Firstly, while the material for construction is waterproof, we did not take into account the structural design. In the future, we can further waterproof the system by completely insulating the PCB and all electrical connectors with silicone gel during fabrication. Secondly, the current design carries all IMUs, PCBs, and batteries separately on the elastic sole. Such design of the external soles might be problematic, because of their thickness in some applications. For example, in professional dance learning, dance shoes typically have very thin leather soles. Although the sole might be comfortable enough for normal dancing, it probably requires a modification for the professional dance movement. Future prototypes can miniaturize the whole design by integrating all components into one button-like gadget for more versatile and adaptive applications.
Power Management: Since the MEMS IMUs we used do not consume a lot of power, we can use energy harvesting methods such as near-field communication (NFC) coil to power the electronics [1]. NFC is a wireless technology capable of transmitting power between integrated components through inductive coupling between coils. The transmitter generates an oscillating electromagnetic field, which is then picked up by the receiver to obtain power. Some low-power NFC devices known as NFC tags can be battery-free and economical, yet only capable of working over short distances. Since NFC devices are also typically made of silicone [2], we can seamlessly integrate the NFC coil into the substrate itself as part of the system by using the same material, increasing the overall flexibility and robustness of the wireless system. When combined with an energy collector, the NFC power supply system can capture data when the label is not powered on. Furthermore, we can also implement other methods of battery-free communication and energy collection, such as the ambient backscatter technology [3], the piezoelectric method to collect energy from vibration [4], or the thermoelectric effect to collect energy from heat transfer [5].
Activation Trigger and Auto Segmentation: We envision social foot-to-foot input as a complementary and alternative input method in a variety of future collaborative computing device ecosystems. In real life, the system needs to distinguish between intentional interactions with the device and unintentional daily activities, including walking, running, and so on. To avoid accidental triggering, it can only respond after activation. Users can choose between an "activation pose" and a "termination pose" that contain a special command (or series of commands). In addition, sometimes participants had to enter multiple gestures in sequence in the same context. Currently, our gesture segmentation is based on the difference between resting and moving, but in practice, the user may perform gestures continuously. In this case, we plan on using signal processing or machine learning models for gesture detection and partition to separate users' gestures. Further processing and updates of these algorithms will enable the use of Shoes++ in a real-world setting.
Opportunities for Scalability
Create Gesture Vocabulary with More People: Currently, due to the limited number of devices and capacity of the experiments, we only designed social foot-to-foot gestures based on two and three participants as a trial. We believe the concept of interactive foot gesturing, with a goal to complete tasks collaboratively, is valid as we move into larger groups of participants. Each additional participant will bring an extra layer of complexity when analyzing gesture semantics, and we plan to include more multi-user foot gestures to complete the vocabulary. Moreover, the spatial variations and user experience in larger group-based interactions can also be further explored.
Front-end Software for Shoes++ Design: In our post-study interviews, many participants expressed an interest in the aesthetic customization of Shoes++. They mentioned that they would be willing to wear the device in their daily lives if they have the opportunity to choose from a variety of styles, patterns, materials, and colors. Currently, our Shoes++ is more focused on structural versatility, but in the future, we can improve the appearance design. For example, we can provide a front-end software and appearance rendering tool to help users design their own Shoes++.
Using Synthetic Data to Enlarge Training Set: In real-world applications, collecting training data from users, especially if it is collected from multiple users, is expensive and time-consuming. To solve this problem, we plan to use a synthetic non-word approach to extend the training set. This approach requires a 2D/3D human body gestures simulation system [6] that continuously generates human body models capable of making dynamic foot gestures. We can choose the location of the IMUs on the foot of the 3D human body and calculate the data of virtual IMUs from motion data. A series of works in computer vision has demonstrated the feasibility of controlling 2D/3D human bodies and using virtual IMUs. For example, Kwon et al. proposed IMUTube, which retrieves videos from public repositories and subsequently generates virtual IMU data from the videos [7]. Through these studies, we believe it may be possible to synthesize datasets of various foot movements in multiple people to facilitate future training.
Challenges and Opportunities for Practical Uses
Practical Design Consideration: For a sole to be practically usable, the current design needs to be improved by waterproofing the electronic components and reducing the overall volume of the system. Firstly, while the material for construction is waterproof, we did not take into account the structural design. In the future, we can further waterproof the system by completely insulating the PCB and all electrical connectors with silicone gel during fabrication. Secondly, the current design carries all IMUs, PCBs, and batteries separately on the elastic sole. Such design of the external soles might be problematic, because of their thickness in some applications. For example, in professional dance learning, dance shoes typically have very thin leather soles. Although the sole might be comfortable enough for normal dancing, it probably requires a modification for the professional dance movement. Future prototypes can miniaturize the whole design by integrating all components into one button-like gadget for more versatile and adaptive applications.
Power Management: Since the MEMS IMUs we used do not consume a lot of power, we can use energy harvesting methods such as near-field communication (NFC) coil to power the electronics [1]. NFC is a wireless technology capable of transmitting power between integrated components through inductive coupling between coils. The transmitter generates an oscillating electromagnetic field, which is then picked up by the receiver to obtain power. Some low-power NFC devices known as NFC tags can be battery-free and economical, yet only capable of working over short distances. Since NFC devices are also typically made of silicone [2], we can seamlessly integrate the NFC coil into the substrate itself as part of the system by using the same material, increasing the overall flexibility and robustness of the wireless system. When combined with an energy collector, the NFC power supply system can capture data when the label is not powered on. Furthermore, we can also implement other methods of battery-free communication and energy collection, such as the ambient backscatter technology [3], the piezoelectric method to collect energy from vibration [4], or the thermoelectric effect to collect energy from heat transfer [5].
Activation Trigger and Auto Segmentation: We envision social foot-to-foot input as a complementary and alternative input method in a variety of future collaborative computing device ecosystems. In real life, the system needs to distinguish between intentional interactions with the device and unintentional daily activities, including walking, running, and so on. To avoid accidental triggering, it can only respond after activation. Users can choose between an "activation pose" and a "termination pose" that contain a special command (or series of commands). In addition, sometimes participants had to enter multiple gestures in sequence in the same context. Currently, our gesture segmentation is based on the difference between resting and moving, but in practice, the user may perform gestures continuously. In this case, we plan on using signal processing or machine learning models for gesture detection and partition to separate users' gestures. Further processing and updates of these algorithms will enable the use of Shoes++ in a real-world setting.
Opportunities for Scalability
Create Gesture Vocabulary with More People: Currently, due to the limited number of devices and capacity of the experiments, we only designed social foot-to-foot gestures based on two and three participants as a trial. We believe the concept of interactive foot gesturing, with a goal to complete tasks collaboratively, is valid as we move into larger groups of participants. Each additional participant will bring an extra layer of complexity when analyzing gesture semantics, and we plan to include more multi-user foot gestures to complete the vocabulary. Moreover, the spatial variations and user experience in larger group-based interactions can also be further explored.
Front-end Software for Shoes++ Design: In our post-study interviews, many participants expressed an interest in the aesthetic customization of Shoes++. They mentioned that they would be willing to wear the device in their daily lives if they have the opportunity to choose from a variety of styles, patterns, materials, and colors. Currently, our Shoes++ is more focused on structural versatility, but in the future, we can improve the appearance design. For example, we can provide a front-end software and appearance rendering tool to help users design their own Shoes++.
Using Synthetic Data to Enlarge Training Set: In real-world applications, collecting training data from users, especially if it is collected from multiple users, is expensive and time-consuming. To solve this problem, we plan to use a synthetic non-word approach to extend the training set. This approach requires a 2D/3D human body gestures simulation system [6] that continuously generates human body models capable of making dynamic foot gestures. We can choose the location of the IMUs on the foot of the 3D human body and calculate the data of virtual IMUs from motion data. A series of works in computer vision has demonstrated the feasibility of controlling 2D/3D human bodies and using virtual IMUs. For example, Kwon et al. proposed IMUTube, which retrieves videos from public repositories and subsequently generates virtual IMU data from the videos [7]. Through these studies, we believe it may be possible to synthesize datasets of various foot movements in multiple people to facilitate future training.
Overview of Shoes++. (a) IMU-mounted, detachable smart soles; (b) Smart soles in use as two users perform social foot-to-foot interaction; (c) Part of the input vocabulary of social foot-to-foot gestures.
Four shoe sole prototypes in our iterative design include (a) the first version, in which two layers separate easily once the electronic components are placed inside; (b) the second version with a thick and bulky bottom that can hardly stretch to fit on shoes; (c) the third version, which has room for improvement in flexibility and aesthetics; (d) the final version. The front (e) and back (f) of the final version show an optimized design to increase stretchability and house all electronic components, which are shown in (g).
Examples of two-user interaction based on foot gestures. (a) Creative Collaboration: different objects are created in AR when the users pose different sets of gestures. (b) Collaboration Guidance: Shoes++ can give real-time feedback on timing and foot movement in social dance learning. (c) Emotion Enhancement: instead of playing an arcade dance game on their own, players collaboratively perform dance moves to finish the game.
Three-person foot-to-foot gestures. Five motion types (rows) are divided into synchronous and asynchronous in timing. The motion type Jump with Same Direction (JumpS for short) and Jump with Different Directions (JumpD for short), i.e. the first and second row in the figure, are used to describe the moving direction of two feet of each person. Five columns indicate their positional change: gather (three people get closer when doing the gesture), spread (three people get far away when doing the gesture), clockwise, counterclockwise, and (each person’s center of gravity) position (remains) unchanged. Different colors are used to mark different people.
Overview of Shoes++. (a) IMU-mounted, detachable smart soles; (b) Smart soles in use as two users perform social foot-to-foot interaction; (c) Part of the input vocabulary of social foot-to-foot gestures.
Four shoe sole prototypes in our iterative design include (a) the first version, in which two layers separate easily once the electronic components are placed inside; (b) the second version with a thick and bulky bottom that can hardly stretch to fit on shoes; (c) the third version, which has room for improvement in flexibility and aesthetics; (d) the final version. The front (e) and back (f) of the final version show an optimized design to increase stretchability and house all electronic components, which are shown in (g).
Examples of two-user interaction based on foot gestures. (a) Creative Collaboration: different objects are created in AR when the users pose different sets of gestures. (b) Collaboration Guidance: Shoes++ can give real-time feedback on timing and foot movement in social dance learning. (c) Emotion Enhancement: instead of playing an arcade dance game on their own, players collaboratively perform dance moves to finish the game.
Three-person foot-to-foot gestures. Five motion types (rows) are divided into synchronous and asynchronous in timing. The motion type Jump with Same Direction (JumpS for short) and Jump with Different Directions (JumpD for short), i.e. the first and second row in the figure, are used to describe the moving direction of two feet of each person. Five columns indicate their positional change: gather (three people get closer when doing the gesture), spread (three people get far away when doing the gesture), clockwise, counterclockwise, and (each person’s center of gravity) position (remains) unchanged. Different colors are used to mark different people.
Overview of Shoes++. (a) IMU-mounted, detachable smart soles; (b) Smart soles in use as two users perform social foot-to-foot interaction; (c) Part of the input vocabulary of social foot-to-foot gestures.
Four shoe sole prototypes in our iterative design include (a) the first version, in which two layers separate easily once the electronic components are placed inside; (b) the second version with a thick and bulky bottom that can hardly stretch to fit on shoes; (c) the third version, which has room for improvement in flexibility and aesthetics; (d) the final version. The front (e) and back (f) of the final version show an optimized design to increase stretchability and house all electronic components, which are shown in (g).
Examples of two-user interaction based on foot gestures. (a) Creative Collaboration: different objects are created in AR when the users pose different sets of gestures. (b) Collaboration Guidance: Shoes++ can give real-time feedback on timing and foot movement in social dance learning. (c) Emotion Enhancement: instead of playing an arcade dance game on their own, players collaboratively perform dance moves to finish the game.
Three-person foot-to-foot gestures. Five motion types (rows) are divided into synchronous and asynchronous in timing. The motion type Jump with Same Direction (JumpS for short) and Jump with Different Directions (JumpD for short), i.e. the first and second row in the figure, are used to describe the moving direction of two feet of each person. Five columns indicate their positional change: gather (three people get closer when doing the gesture), spread (three people get far away when doing the gesture), clockwise, counterclockwise, and (each person’s center of gravity) position (remains) unchanged. Different colors are used to mark different people.
[1] Madjda Bouklachi, Marc Biancheri-Astier, Antoine Diet, and Yann Le Bihan. 2019. Energy Harvesting of a NFC Flexible Patch for Medical Applications. In 2019 IEEE Wireless Power Transfer Conference (WPTC). IEEE, 2019 IEEE Wireless Power Transfer Conference, 249–252.
[2] Justin Chan and Shyamnath Gollakota. 2017. Data storage and interaction using magnetized fabric. In Proceedings of the 30th Annual ACM Symposium on User Interface Software and Technology. Association for Computing Machinery, 655–663.
[3] Nivedita Arora and Gregory D Abowd. 2018. ZEUSSS: Zero energy ubiquitous sound sensing surface leveraging triboelectric nanogener-ator and analog backscatter communication. In The 31st Annual ACM Symposium on User Interface Software and Technology Adjunct Proceedings. Association for Computing Machinery, 81–83.
[4] Heung Soo Kim, Joo-Hyong Kim, and Jaehwan Kim. 2011. A review of piezoelectric energy harvesting based on vibration. International journal of precision engineering and manufacturing 12, 6 (2011), 1129–1141.
[5] Amin Nozariasbmarz, Henry Collins, Kelvin Dsouza, Mobarak Hossain Polash, Mahshid Hosseini, Melissa Hyland, Jie Liu, Abhishek Malhotra, Francisco Matos Ortiz, Farzad Mohaddes, Viswanath Padmanabhan Ramesh, Yasaman Sargolzaeiaval, Nicholas Snouwaert, Mehmet C. Özturk, and Daryoosh Vashaee. 2020. Review of wearable thermoelectric energy harvesting: From body temperature to electronic systems. Applied Energy 258 (2020), 114069.
[6] Ian Gibson, Bhat Nikhil Jagdish, Gao Zhan, Khatereh Hajizedah, Hyunh Kim Tho, Huang Mengjie, and Chevanthie Dissanayake. 2012. Development of a human spine simulation system. Advances in Therapeutic Engineering 0, 0 (2012), 1.
[7] Hyeokhyen Kwon, Bingyao Wang, Gregory D Abowd, and Thomas Plötz. 2021. Approaching the Real-World: Supporting Activity Recognition Training with Virtual IMU Data. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies 5, 3 (2021), 1–32.
[1] Madjda Bouklachi, Marc Biancheri-Astier, Antoine Diet, and Yann Le Bihan. 2019. Energy Harvesting of a NFC Flexible Patch for Medical Applications. In 2019 IEEE Wireless Power Transfer Conference (WPTC). IEEE, 2019 IEEE Wireless Power Transfer Conference, 249–252.
[2] Justin Chan and Shyamnath Gollakota. 2017. Data storage and interaction using magnetized fabric. In Proceedings of the 30th Annual ACM Symposium on User Interface Software and Technology. Association for Computing Machinery, 655–663.
[3] Nivedita Arora and Gregory D Abowd. 2018. ZEUSSS: Zero energy ubiquitous sound sensing surface leveraging triboelectric nanogener-ator and analog backscatter communication. In The 31st Annual ACM Symposium on User Interface Software and Technology Adjunct Proceedings. Association for Computing Machinery, 81–83.
[4] Heung Soo Kim, Joo-Hyong Kim, and Jaehwan Kim. 2011. A review of piezoelectric energy harvesting based on vibration. International journal of precision engineering and manufacturing 12, 6 (2011), 1129–1141.
[5] Amin Nozariasbmarz, Henry Collins, Kelvin Dsouza, Mobarak Hossain Polash, Mahshid Hosseini, Melissa Hyland, Jie Liu, Abhishek Malhotra, Francisco Matos Ortiz, Farzad Mohaddes, Viswanath Padmanabhan Ramesh, Yasaman Sargolzaeiaval, Nicholas Snouwaert, Mehmet C. Özturk, and Daryoosh Vashaee. 2020. Review of wearable thermoelectric energy harvesting: From body temperature to electronic systems. Applied Energy 258 (2020), 114069.
[6] Ian Gibson, Bhat Nikhil Jagdish, Gao Zhan, Khatereh Hajizedah, Hyunh Kim Tho, Huang Mengjie, and Chevanthie Dissanayake. 2012. Development of a human spine simulation system. Advances in Therapeutic Engineering 0, 0 (2012), 1.
[7] Hyeokhyen Kwon, Bingyao Wang, Gregory D Abowd, and Thomas Plötz. 2021. Approaching the Real-World: Supporting Activity Recognition Training with Virtual IMU Data. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies 5, 3 (2021), 1–32.
[1] Madjda Bouklachi, Marc Biancheri-Astier, Antoine Diet, and Yann Le Bihan. 2019. Energy Harvesting of a NFC Flexible Patch for Medical Applications. In 2019 IEEE Wireless Power Transfer Conference (WPTC). IEEE, 2019 IEEE Wireless Power Transfer Conference, 249–252.
[2] Justin Chan and Shyamnath Gollakota. 2017. Data storage and interaction using magnetized fabric. In Proceedings of the 30th Annual ACM Symposium on User Interface Software and Technology. Association for Computing Machinery, 655–663.
[3] Nivedita Arora and Gregory D Abowd. 2018. ZEUSSS: Zero energy ubiquitous sound sensing surface leveraging triboelectric nanogener-ator and analog backscatter communication. In The 31st Annual ACM Symposium on User Interface Software and Technology Adjunct Proceedings. Association for Computing Machinery, 81–83.
[4] Heung Soo Kim, Joo-Hyong Kim, and Jaehwan Kim. 2011. A review of piezoelectric energy harvesting based on vibration. International journal of precision engineering and manufacturing 12, 6 (2011), 1129–1141.
[5] Amin Nozariasbmarz, Henry Collins, Kelvin Dsouza, Mobarak Hossain Polash, Mahshid Hosseini, Melissa Hyland, Jie Liu, Abhishek Malhotra, Francisco Matos Ortiz, Farzad Mohaddes, Viswanath Padmanabhan Ramesh, Yasaman Sargolzaeiaval, Nicholas Snouwaert, Mehmet C. Özturk, and Daryoosh Vashaee. 2020. Review of wearable thermoelectric energy harvesting: From body temperature to electronic systems. Applied Energy 258 (2020), 114069.
[6] Ian Gibson, Bhat Nikhil Jagdish, Gao Zhan, Khatereh Hajizedah, Hyunh Kim Tho, Huang Mengjie, and Chevanthie Dissanayake. 2012. Development of a human spine simulation system. Advances in Therapeutic Engineering 0, 0 (2012), 1.
[7] Hyeokhyen Kwon, Bingyao Wang, Gregory D Abowd, and Thomas Plötz. 2021. Approaching the Real-World: Supporting Activity Recognition Training with Virtual IMU Data. Proceedings of the ACM on Interactive, Mobile, Wearable and Ubiquitous Technologies 5, 3 (2021), 1–32.
Updated on 12/28/2023, made with ❤ by Jiayi Zhou